General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

Users may download and print one copy of any publication from the public portal for the purpose of private study or research.

You may not further distribute the material or use it for any profit-making activity or commercial gain

You may freely distribute the URL identifying the publication in the public portal If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from orbit.dtu.dk on: Feb 04, 2020

Revealing the compact structure of lactic acid bacterial hetero-exopolysaccharides bySAXS and DLS

Khan, Sanaullah; Birch, Johnny; Harris, Pernille; Van Calsteren, Marie-Rose; Ipsen, Richard; Peters,Günther H.J.; Svensson, Birte; Almdal, KristofferPublished in:Biomacromolecules

Link to article, DOI:10.1021/acs.biomac.6b01597

Publication date:2017

Document VersionPeer reviewed version

Link back to DTU Orbit

Citation (APA):Khan, S., Birch, J., Harris, P., Van Calsteren, M-R., Ipsen, R., Peters, G. H. J., ... Almdal, K. (2017). Revealingthe compact structure of lactic acid bacterial hetero-exopolysaccharides by SAXS and DLS. Biomacromolecules,18(3), 747-757. https://doi.org/10.1021/acs.biomac.6b01597

1

Revealing the compact structure of lactic acid bacterial

hetero-exopolysaccharides by SAXS and DLS

Sanaullah Khan†, Johnny Birch‡, Pernille Harris§, Marie-Rose Van Calsteren∥, Richard Ipsen ,

Günther H.J. Peters§, Birte Svensson‡ and Kristoffer Almdal†*

†Department of Micro- and Nanotechnology, DTU, Ørsteds Plads, building 423, DK-2800 Kgs.

Lyngby, Denmark.

‡Enzyme and Protein Chemistry, Department of Systems Biology, DTU, Elektrovej, building 375, DK-

2800 Kgs. Lyngby, Denmark.

§Department of Chemistry, DTU, Kemitorvet, building 207, DK-2800 Kgs. Lyngby, Denmark

∥Saint-Hyacinthe Research and Development Centre, Agriculture and Agri-Food Canada, 3600

Casavant Boulevard West, Saint-Hyacinthe, Quebec J2S 8E3, Canada.

Department of Food Science, University of Copenhagen, Rolighedsvej 26, DK-1958 Frederiksberg

Denmark

2

ABSTRACT

Molecular structures of exopolysaccharides are required to understand their functions and the

relationships between the structure and physical and rheological properties. Small-angle X-ray

scattering and dynamic light scattering were used in conjunction with molecular modeling to

characterize solution structures of three lactic acid bacterial hetero-exopolysaccharides (HePS-1, HePS-

2 and HePS-3). Values of radius of gyration RG, cross-sectional radius of gyration RXS, approximate

length L and hydrodynamic diameter were not directly proportional to the molar mass and indicated the

HePSs adopted a compact coil-like rather than an extended conformation. Constrained molecular

modeling of 15,000 randomised HePS-1 conformers resulted in five best-fit structures with R factor

of 3.94.6% revealing random coil-like structure. Φ and Ψ angle analysis of glycosidic linkages

in HePS-1 structures suggests Galf residues significantly influence the conformation. Ab initio

scattering modeling of HePS-2 and HePS-3 gave excellent curve fittings with χ2 of 0.43 and 0.34 for

best-fit models, respectively, compatible with coil-like conformation. The findings disclose solution

behaviour of HePS relevant for their interactions with biomacromolecules e.g. milk proteins.

Keywords: Hetero-exopolysaccharides; small-angle X-ray scattering; constrained scattering modeling;

dynamic light scattering

3

INTRODUCTION

Exopolysaccharides (EPSs) are biopolymers secreted by a vast number of microorganisms including

yeasts, molds, microalgae and bacteria.1,2 EPSs have wide diversity in monosaccharide composition

and glycosidic linkages of their oligosaccharide repeat units.3 Bacterial EPSs are grouped into two

classes, namely homoexopolysaccharides (HoPSs) and heteroexopolysaccharides (HePSs). HoPSs

contain a single type of monosaccharide units (mostly D-glucopyranose or D-fructofuranose) and are

synthesized from sucrose by glucansucrases and fructansucrases, respectively. 4 HePSs by contrast are

composed of more than one type of monosaccharide in oligosaccharide repeats of 3 to 9 residues

synthesized from intracellular sugar nucleotide precursors and then polymerized.58 Typical lactic acid

bacteria (LAB) HoPS and HePS contain a linear polysaccharide backbone with branches of varying

lengths.

The physiological function of EPSs is not fully known. Generally, in their natural environment EPSs

are believed to protect microbial cells against desiccation, phagocytosis, antibiotics, environmental

stress and are involved in adhesion to solid surfaces, biofilm formation and cell recognition.710 EPSs

receive considerable attention in various food and pharmaceutical industries due to their technological

properties and GRAS (”generally regarded as safe”) status.11, 12 In the food industry, LAB-EPSs serve

as viscosifying, emulsifying and gelling agents to improve rheological and textural properties of food

products.9, 12, 13 Moreover, LAB-EPSs have also been described to exert health benefits, e.g. lowering

cholesterol levels,10, 14, 15 reducing formation of biofilms,16 modulating adhesion to epithelial cells and

inducing a prebiotic effect by increasing bifidobacteria levels in the intestinal tract.17, 18

The efficacious applications of EPSs are considered to result from their capability to regulate

viscosity and bind water, which correlate to their structure and interactions with proteins.17, 19, 20 The

4

molecular weight of EPSs is one of several factors contributing to their functional properties in

solution.21 Other factors including the three-dimensional (3D) structure and conformational flexibility

connected with degree of branching,22, 23 charge,24, 25 and ability to engage in intermolecular association

significantly influence physical and rheological properties of EPSs.10 The nature of glycosidic linkages

and sugar ring configuration also impacts viscosity. High viscosities are associated with high molar

mass, chain stiffness and extended structures of EPSs which tend to occupy larger volumes in

solution.21, 26 The degree of chain stiffness is related to the presence of glycosidic linkages with less

rotational freedom in the EPS structure. For example β-(1→4) linkages are stiffer than β-(1→3) and β-

(1→2) glycosidic linkages, hence contributing to the viscosity.27

To expand knowledge on EPS conformation in solution and the relationship to physical and

rheological properties, experimental analysis is needed of the molecular structures of intact EPSs. High

resolution X-ray crystallography and nuclear magnetic resonance (NMR) methods that are widely used

to obtain 3D structures of biomacromolecules are not suited for EPS structure determination as X-ray

crystallography requires well ordered crystals of good diffraction quality, and NMR is limited to

comparatively small macromolecules. The extreme flexibility, large size, high viscosity and modest

water solubility of EPSs precluded their 3D structure determination to date. In the present study, we

characterize three HePSs (HePS-1, HePS-2 and HePS-3) from Streptococcus thermophilus28,

Lactobacillus rhamnosus29 and Lactobacillus delbrueckii ssp. bulgaricus 30 using small-angle X-ray

scattering (SAXS) and Dynamic Light Scattering (DLS). The solution scattering in conjunction with

constrained scattering modeling was used to determine the molecular structure of HePS-1. This

approach is well-established and has been used for solution structure determination of multi-domain

proteins and anionic oligosaccharides,3134 but has not previously been used for structural analysis of

5

large intact EPSs. The data provide new insight into structural features of HePSs and form a basis for

identifying structural requirements enabling binding to biomacromolecules such as milk proteins.

MATERIALS AND METHODS

Production, purification and quantification of HePSs. HePS-1, HePS-2 and HePS-3 were

produced in 10% skimmed milk medium (10% skimmed milk powder in MilliQ (MQ) water) by

Streptococcus thermophilus EU20 (kind gift of Prof. Luc De Vuyst, Vrije Universiteit Brussels,

Belgium),28 Lactobacillus rhamnosus GG (ATCC 53103)29 and Lactobacillus delbrueckii ssp.

bulgaricus (NCIMB 702483)30, respectively.

HePS-1 was isolated from the culture medium as described previously.35 Briefly, proteins and cells

were precipitated by addition of trichloroacetic acid (TCA) to a final concentration of 20% and

removed by centrifugation (15,000 g, 30 min, 4ºC). HePS-1 was precipitated overnight (5ºC) from the

supernatant by adding ethanol (2:1, v/v), isolated by centrifugation as above, redissolved in MQ water

(1 h, 50ºC) and residual proteins were precipitated by TCA followed by ethanol-precipitation of HePS-

1 (both steps as above). HePS-1 was dissolved in MQ-water as above, dialyzed (50,000 kDa cutoff,

20ºC) against MQ water (5 d, water replacement twice a day) and finally purified by size-exclusion

chromatography (XK26/100 Sephacryl S-400; GE Healthcare, Uppsala, Sweden) and lyophilized.

HePS-2 and HePS-3 were purified from cultures by TCA precipitation (final concentration 20% (w/v))

of cells and proteins for 20 min at room temperature, centrifugation (as above), followed by the HePS

precipitation overnight by adding acetone (1:1 v/v; 4°C), centrifugation (12,000 g, 30 min, 4ºC),

redissolution in MQ water, extraction by 1:1 (v/v) phenol/chloroform/isoamyl alcohol 25:24:1 twice and

extraction by 1:1 (v/v) of chloroform/isoamyl alcohol 24:1 once

6

. Following acetone precipitation and centrifugation (as above), the purified HePS was dissolved in

deionized water, dialyzed, filtered and freeze-dried.

HePS concentrations were quantified using the phenol–sulfuric acid method and monosaccharide

mixtures corresponding to the composition of the various repeat units as standards.36

Molecular weight (Mw) Determination. Molecular weights of HePS-1 HePS-3 were determined

by size exclusion chromatography (SEC). The SEC system consists of a solvent delivery system (LC-

10AD), autosampler (SIL-10A), RI detector (RID-10A; all Shimadzu, Kyoto, Japan) and a size

exclusion column (OH-PK SB-805 HQ 300 × 8 mm, Tokyo, Japan). Dextran standards of Mw 4.5 MDa,

1.45 MDa, 560 kDa, 350 kDa (American Polymer Standards Corporation, Mentor, Ohio, USA), 276.5

kDa, 196.3 kDa, 123 kDa, 43 kDa (Pharmacosmos, Holbaek, Denmark) and pullulan of 22 kDa were

used for calibration. Standards (1–2.7 mg ml−1) and HePSs (1–2 mg ml−1) were dissolved in mobile

phase (10 mM sodium citrate/citric acid pH 4.0), degassed, kept overnight, filtered (0.45 µm filters;

Frisenette ApS, Knebel, Denmark) and 100 µl was subjected to SEC at a flow rate of 0.5 ml min−1.

NMR characterization of HePS-3. HePS-3 (~10 mg/mL) was exchanged in D2O, and the NMR

spectrum was obtained on a Chemagnetics (Fort Collins, CO, USA) CMX Infinity 300 spectrometer

equipped with a 5-mm dual 13C/1H Nalorac (Martinez, CA, USA) probe. The temperature was

controlled at 70°C. The 1H NMR experiment was performed with original pulse program of the

Spinsight software. For the 1D spectrum, 8 K complex data points were acquired and processed by

exponential multiplication with a line broadening factor equal to the digital resolution, complex Fourier

transform, phase correction and third-order polynomial baseline correction.

Small angle X-ray scattering (SAXS). Experiments with HePS-1, HePS-2 and HePS-3 were

performed on beamline I911-4 at the MAX IV (synchrotron radiation facility, Lund Sweden).37 The

7

sample detector distance and the direct beam position were calibrated using silver behenate. Pure water

was used to place the data on an absolute scale. Buffers were measured before each sample and

averaged before subtraction. The sample size was approximately 40 µL using an automated flow cell.

Eight time frames were measured using 20 s exposure time optimized using on-line checks during

acquisitions to avoid radiation damage effects. The frames were averaged to maximize signal-to-noise

ratios. Initial data treatment was done using the in-house software Bli911-4. Buffer averaging and

subtraction prior to data analysis was done in Primus.38 Data were collected for all three HePSs at

concentrations of 12 mg/mL in 10 mM sodium citrate, pH 4.0. Guinier analysis gives the radius of

gyration RG, which measures the degree of structural elongation in solution if the internal

inhomogeneity of scattering within the macromolecules has no effect. Guinier plots at low Q (where Q

= 4 π sin θ/λ; 2θ is the scattering angle and λ is the wavelength) give RG and the forward scattering at

zero angle I(0):39

ln I(Q) = ln I(0) - RG2Q2/3

If the structure is elongated (i.e. rod-shaped), the radius of gyration of the cross-sectional structure

RXS and the mean cross-sectional intensity at zero angle [I(Q).Q]Q0 parameters are obtained from:

ln [I(Q).Q] = ln [I(Q).Q]Q 0 - RXS2Q2/2

The RG and RXS analyses were done using PRIMUS.38 Indirect Fourier transformation of the full

scattering curve I(Q) in reciprocal space gives the distance distribution function P(r) in real space. This

yields the maximum dimension of the macromolecule L and its most commonly occurring distance

vector M in real space:

8

QQrQrQIrP d )sin( )( 2π21

= )(o

The transformation was carried out using GNOM software.40

Molecular modeling of HePS-1. 3D models of HePS-1HePS-3 repeating units were built using the

Sweet2 web server (sweet2)41, 42 from their structures determined by NMR (Figure 1). The Φ and Ψ

angle values in HePS-1 repeating unit are shown in Table 2. A fully extended HePS-1 structure was

generated by joining 128 repeat units through glycosidic linkages using Discovery Studio (version 4.5)

molecular graphics software (Accelrys, San Diego, CA). Fully extended models for HePS-2 and HePS-

3 were also generated by joining about 389 and 206 repeating units, respectively using Discovery

Studio. These models were generated in accordance with their molecular weights determined by size

exclusion chromatography.

A total of 15,000 conformationally randomized models for HePS-1 were generated from the

extended (linear) model using the TorsionKick function in a PERL script modified from the

ExtractAngle.pl script provided with the Discovery Studio software.33, 34 The TorsionKick takes any

value in a maximum range of ±45° from their starting values. A constant force field termed

DREIDING minimization was used to avoid steric clashes between the atoms in each randomized

structure.43

Constrained scattering modeling of HePS-1. X-ray scattering curves were calculated for each

randomized HePS-1 model for comparison with the experimental curves using the SCT program.44, 45 A

cube side length of 0.527 nm and a cutoff of 4 atoms were used to create spheres in the HePS-1 models.

The hydration shell (0.3 g H2O/ g protein) was created by adding spheres to the unhydrated sphere

9

models using HYPRO.45, 46 The X-ray scattering curve I(Q) was calculated using the Debye equation as

adapted to spheres and assuming a uniform scattering density for the spheres: 44,47

) sin

1=

2-2 + 1-( )( = (0)

)(

Qr j

Qr jA j

m

jnnQg

I

QI

RQQRQRQRQg 66 /)2) cos - sin(3( = )(

where g(Q) is the squared form factor for the sphere of radius r, n is the number of spheres filling the

body, Aj is the number of distances rj for that value of j, rj is the distance between the spheres, and m is

the number of different distances rj. The models were assessed by calculation of the radius of gyration

RG and cross-sectional RXS values in the same Q ranges used for the experimental Guinier fits. Models

that were filtered using the RG and RXS filters were then ranked using a goodness-of-fit R factor in order

to identify the best-fit models.

Ab initio scattering modeling of HePS-2 and HePS-3. SAXS data of HePS-2 and HePS-3 were

used to derive ab initio information on their shapes and conformation. To generate a molecular

envelope of HePS-2 and HePS-3 from scattering data, 20 ab initio beads models were generated for

each HePS using DAMMIF program,48 assuming no P1 symmetry. These models were filtered and

averaged in DAMAVER47 in the ATSAS package.50

Dynamic light scattering (DLS). Particle size distribution data for HePS-1, HePS-2 and HePS-3

were obtained at a scattering angle of 90° (BI-200SM instrument; Brookhaven Instruments

Corporation; Holtsville, NY, USA) and 23°C. DLS analyses the size distribution of particles

undergoing Brownian motion through illumination by a laser beam. The time-dependent fluctuations in

the intensity of scattered light reflect the rate at which the particles diffuse. The distributions of mean

apparent translational diffusion coefficients (DT) were determined by fitting the DLS autocorrelation

10

functions obtained by the Brookhaven system using non-negative constrained least-squares (NNLS).

The distribution of DT values was converted to the distribution of hydrodynamic diameters (dH) using

the Stokes-Einstein equation:

dH = kT/3πηDT

where k is the Boltzmann constant, T the absolute temperature, and η the solvent viscosity (assumed to

be that of water, 0.933 mPa·s).

RESULTS AND DISCUSSION

HePSs repeat units and molecular weight (Mw). Primary structures of HePS-1, HePS-2 andHePS-

3 consist of a branched repeat units of dp7 (where dp stands for degree of polymerization) composed

of glucose, rhamnose and galactose (molar ratio of 2:2:3),28 dp6 of N-acetyl-glucosamine, rhamnose

and galactose (molar ratio of 1:1:4),29 dp7 of glucose, rhamnose and galactose (molar ratio of 1:1:5) as

shown in Figure 1. Analysis of the 1H NMR spectrum of HePS-3 (data not shown) revealed that the

HePS-3 repeating unit structure is identical to that previously described for the HePS purified from

Lactobacillus delbrückii subspecies bulgaricus strain rr.35 The Mw was determined to be 145 ± 4 kDa

for HePS-1, 389 ± 6 kDa for HePS-2 and 372 ± 8 kDa for HePS-3 by SEC (data not shown).

11

Figure 1. Primary structures of HePS-1, HePS-2 andHePS-3 repeating units. HePS-1 was purified from Streptococcus thermophilus EU20, HePS-2 from Lactobacillus rhamnosus GG (ATCC 53103) and HePS-3 from Lactobacillus delbrueckii ssp. bulgaricus (NCIMB 702483).

SAXS of HePS-1, HePS-2 andHePS-3. HePS-1,HePS-2 and HePS-3 (12 mg/mL) were

characterized by SAXS to gain insight into their overall structure in solution.31 The Guinier

analyses of lnI(Q) versus Q2 at low Q values give the radius of gyration RG that monitors the extent

of macromolecular elongation. The Guinier fits resulted in an RG value of 10.14 ± 0.31 nm for

HePS-1 using a Q range of 0.1–0.15 nm−1, one of 11.72 ± 0.23 nm for HePS-2 using a Q range of

0.09–0.12 nm−1 and one of 10.82 ± 0.48 nm for HePS-3 using a Q range of 0.09–0.13 nm−1 (Figure

2A, Table 1). The experimental RG values do not show a linear relationship with modeled RG values

of 91 nm, 230 nm and 139 nm calculated for extended chain lengths of HePS-1, HePS-2 and HePS-

3, respectively, indicating that the HePSs are not completely extended rods but adopt coil-like

conformations in solution.

For elongated macromolecules, the cross-sectional radius of gyration (RXS) provides

information on the degree of bending within the molecular length. The RXS analyses of

12

lnI(Q)Q versus Q2 showed good fits for HePS-1, HePS-2 and HePS-3 in a Q range of 0.180.31

nm−1, 0.180.27 nm−1 and 0.170.27 nm−1 respectively, resulting in experimental RXS values of

3.22 ± 0.09 nm for HePS-1, 3.95 ± 0.01 nm for HePS-2 and 3.50 ± 0.09 nm for HePS-3 (Figure 2B,

Table 1). If the HePSs are approximated as elliptical cylinders, the combination of RG and RXS

according to the relationship L2 =12(RG2 − RXS

2) resulted in experimental approximate length L of

33.3, 38.2 and 35.5 nm for HePS-1, HePS-2 and HePS-3, respectively.

The distance distribution function P(r) represents the distribution of all the distances

between the atoms within the macromolecule. It provides RG values and model-independent

determinations of overall lengths, L. The mean RG obtained from the P(r) curves were 10.89 ± 0.16

nm, 11.49 ± 0.25 nm and 11.44 ± 0.1 nm for HePS-1, HePS-2 and HePS-3, respectively (Table 1),

which agree with the values obtained from the corresponding Guinier RG analyses.

13

Figure 2. Experimental Guinier RG, RXS and P(r) analyses of HePS-1, HePS-2 and HePS-3. (A) Guinier RG plots for HePS-1 at 1 mg/ml, and HePS-2 and HePS-3 at 2.0 mg/ml. Filled circles represent the Q ranges used to determine the radius of gyration RG based on the best-fit lines as shown. The Q ranges used for RG analyses were 0.1–0.15 nm−1 for HePS-1, 0.09–0.12 nm−1 for HePS-2 and 0.09–0.13 nm−1 for HePS-3. (B) Guinier cross-sectional RXS plots for HePS-1, HePS-2 and HePS-3. The filled circles represent the Q ranges used to determine the RXS values based on the best-fit lines as shown. The Q ranges used for RXS analyses were 0.18–0.31 nm−1 for HePS-1, 0.18–0.27 nm−1 for HePS-2 and 0.17–0.27 nm−1 for HePS-3. (C) Normalized P(r) curve for HePS-

1HePS-2. The M1 value of 8.7 nm and L value of 41.0 nm were observed for HePS-1 (solid line). The M1 value of 9.9 nm and L value of 38.0 nm were observed for HePS-2 (broken line). (D) Normalized P(r) curve for HePS-3. The r values of the maximum at M1 and M2 were 3.9 nm and 11.3 nm, respectively, and the maximum length L values as 41.8.

Values of the maximum length L were determined from the r value where the P(r) curve reaches zero

at large r. The experimental L values were 41.0 nm, 38.0 nm and 41.8 nm for HePS-1, HePS-2 and

HePS-3, respectively (Figure 2C and D). Although the three HePSs differ in molar mass, their lengths

14

Table 1. SAXS and scattering modeling fits for the solution structures of HePSs

HePSs

Filter

Number of models

RG (nm) a

RXS (nm)

R factor (%)

χ2

Length L (nm)

dH (nm)

HePS-1

None

RG, RXS, R factor

Best-fit

Experimental

15,000

5

1

2.80–17.80

9.90–10.84

10.35

10.14 ± 0.31

10.89 ± 0.16

0.04–7.9

3.14–3.47

3.33

3.22 ± 0.09

3.9–32.8

3.9–4.6

3.9

n.a

42–44

42.0

41.0

29.9 ± 0.2

HePS-2

Ab initio models

Best-fit

Experimental

20

1

11.50–11.61

11.56

11.72 ± 0.23 11.49 ± 0.25

3.95 ± 0.01

0.43–0.50

0.43

37–42

39.6

38.0

34.1 ± 0.3

HePS-3 Ab initio models

Best-fit

Experimental

20

1

11.45–11.63

11.63

10.82 ± 0.48 11.44 ± 0.1

3.50 ± 0.09

0.34–0.40

0.34

35.6–41.0

39.0

41.8

35.6 ± 0.5

aThe first value is from the Guinier RG analyses, and the second from the GNOM P(r) analyses.

15

are very close each other, showing that the HePSs of higher molar mass are comparatively more

compact than the one of lower molar mass. These L values are much smaller than those calculated

assuming that the HePS-1, HePS-2 and HePS-3 models have maximally extended molecular

conformations of lengths 321 nm, 790 nm and 475 nm, respectively. This large difference indicated

that the HePSs do not have extended (linear) backbone conformations in solution, being compact

instead. The approximate L values of HePS-1 and HePS-3 estimated from RG and RXS values were

comparable to those from the P(r) curves. The L values of HePS-2 from RG and RXS values and the P(r)

curves were in similar range. It is important to mention that these experimental L values from the P(r)

curves are more accurate than the L values determined from the RG and RXS analyses and the

hydrodynamic diameters measured by DLS (Table 1). Even so, the DLS data and these approximate

lengths show that RG and RXS values are consistent with the P(r) analyses. The maximum M in P(r)

curves corresponds to the most frequently occurring interatomic distance within the HePSs. For

HePS-1 and HePS-2, one peak was observed at r values of 8.7 nm and 9.9 nm respectively

(Figure 2C), while for HePS-3, two maxima M1 and M2 were observed at r values of 3.9 nm and

11.3 nm respectively (Figure 2D).

The HePSs have many similarities to synthetic polymers. The isolated HePS-1, HePS-2 and HePS-3

show a distribution of chain lengths as evidenced by the width of the peaks observed in size exclusion

chromatography (Figure S1). Furthermore the observed compact sizes in solution by SAXS indicate

that the HePSs are relatively flexible chains. For a completely flexible chain in dilute solution only one

length scale is expected, namely the size of the random coil. The fact that we need two length-scales,

RG2 and RXS

2, to describe the elliptical cylinder chain conformation indicates limited flexibility in these

HePSs similar to a description by the worm-like-chain model.51, 52

16

Constrained modeling of HePS-1. Compared to the linear HePS-1 model, the experimental RG,

RXS and maximum length L values showed that HePS-1 is not extended but exhibits a coiled

conformation in solution. Here we used a constrained scattering modeling approach to provide

molecular explanation for these scattering data. Modeling for HePS-1 was performed using the

extended model created using its molar mass and repeat structure as a starting constraint. In the

models, the glycosidic linkage connectivity between oligosaccharide rings was maintained, while the

phi (Φ) and psi (Ψ) angles at each glycosidic linkage (Figure 1) were varied randomly in a range of up

to ±45° from their starting values. A total of 15,000 conformational models for HePS-1 were

generated. Scattering curves for these randomized models were calculated and fitted to the

experimental curves and RG, RXS and R factors were calculated for all modeled curves. The R factor

distributions (Figure 3) showed that all 15,000 HePS-1 randomized models (filled circles) encompassed

the experimental RG and RXS values (dashed lines) and that the best R factor values were well below the

level of 5% that is expected for excellent scattering curve fittings.28 This showed that 15,000

randomized models provided sufficient conformers to be able to determine the best-fit HePS-1

solution structures. The lowest R factor had RG and RXS values (red and blue circles) that agree

well with the experimental RG and RXS values.

The modeling analysis of HePS-1 showed good agreement between models and experimental

scattering data. A family of five best-fit HePS-1 models were selected based on the R factor, RG

and RXS parameters. All five best-fit models gave R factor of 3.9% to 4.6%, RG values of

9.9010.84 nm, RXS values of 3.143.47 nm and maximum length L of 4244 nm (Table 1). All

these agreed well with the corresponding experimental values of 10.14 ± 0.31 nm, 3.22 ± 0.09 nm, and

41.0 nm. In contrast, the linear HePS-1 model gave a much poorer fit with an R factor of 29.9%, RG of

17

Figure 3. Constrained modelling analyses of HePS-1. (A) The R factor values for the 15,000 conformational models of HePS-1 are compared with their RG values. The vertical broken line corresponds to the experimental RG value. The blue circle represents the best-fit model and the other four best-fit models are shown in red close to the R factor minimum. (B) The R factor values for the 15,000 conformational models of HePS-1 are compared with their RXS values. The blue circle represents the best-fit model and the four other best-fit models are shown in red close to the R factor minimum.

91 nm, RXS of 1.1 nm and L of 321 nm. Modeling curve fitting analysis showed excellent visual

agreement between the experimental I(Q) curve and the modeled I(Q) curve, and between

the experimental P(r) curve and the modeled P(r) curve (Figure 4A). In conclusion, the

modeling analyses for the best-fit HePS-1 models show coil-like structures in solution as

may be seen in Figures 4B and D.

18

Figure 4. X-ray constrained scattering modeling curve fits for best-fit and linear-fit HePS-1 models. The data in "A" depicts the I(Q) curves, and the inset shows the P(r) distance distribution functions. The experimental I(Q) and P(r) scattering data are shown by black circles or lines, respectively; the blue line and model correspond to the best-fit HePS-1 model selected from 15,000 conformational models. The green line and model correspond to the linear poor-fit HePS-1 model. The maximum lengths L of best-fit model (blue) is shown in the inset for comparison with their L values in the P(r) curves. (B) The best-fit and linear models are shown in blue and green, respectively (length scales are different). (C) Oligosaccharide fragments generated from the best-fit model in B; the top one shows kink structures that produced by Galf, the bottom one shows a helix-like structure. (D) Superimposition of the five best-fit models for HePS-1. The five best-fit HePS-1 models were superimposed using Discovery Studio VISUALISER software. The best-fit model from Fig. 3 is shown in blue, while the related best-fit models are shown in red.

Conformation of HePS-1. For large, intact HePS, it is not possible to generate

conformationally randomized models and look into their flexibility using NMR and other

techniques. Using a constrained scattering modelling approach, we were able to generate

conformational models and derive structural parameters. In terms of the glycosidic linkages,

information on conformation in the HePS structures was obtained from analyses of the Φ and Ψ

19

angles. The HePSs with stiff glycosidic linkages tend to occupy larger volume in solution

compared to flexible linkages, and hence contribute significantly to increase viscosity. The β-

(1→4) linkages are stiffer than the β-(1→3) and β-(1→2) glycosidic linkages, while the α-linkages

(1→2, 1→3 and 1→6) are flexible.21, 27 Therefore HePSs containing more α-linkages in the

oligosaccharide repeat units are in general more flexible.

The average Φ and Ψ angles of the HeSP-1 α-D-Galp-(1→3)-β-L-Rhap glycosidic bonds

were 76° and 94° (260 values) and β-L-Rhap-(1→4)-β-D-Glcp were 58° and 111° (260 values)

respectively (Table 2). For β-(1→6) linkages (four linkages in total), the average Φ angles ranged

from -86° to 161° (each 260 values) and the Ψ angles ranged from -174° to 159° (each 260

values) (Table 2). Here the standard deviation of Φ angles have increased to 3343°, while the

standard deviation of Ψ angles were similar at 40°. Since the repeating units in HePS-1 contain

three (1→6)-linkages (one α and two β), of which two are consecutive with a five-membered ring

(Galf) which makes HePS-1 more flexible, a very large number of short- and long-range orientations

in the main chain are possible. This could be explained by Φ and Ψ angles analyses and increase

in standard deviation of (1→6) linkages with five-membered ring (Galf), indicating that five-

membered ring Galf residues likely produce kinks or bents in the HePS chain thus showing higher

conformational flexibility (Figure 5). Due to presence of these Galf residues in repeating units

which tend to produce progressive bending or kinks, the HePS-1 chain possesses

secondary structural elements that potentially adopt random coil-like structures in solution (Figure 4B

and C). Modeling and simulation studies (no experimental SAXS data) have also shown that the

HePSs form secondary structural elements, and that the Galf has a significant influence on the

conformational properties of HePSs.53, 54 SAXS studies of HePS (unknown size) from marine

20

bacterium at 0.4% (w/v, i.e. 4 mg/ml) have shown that HePS forms a network of random coiled

polymeric chains that was demonstrated to be relatively homogeneous at high pH and less

homogeneous at very low pH.55 Using the SAXS technique, Stokke et al. have reported that alginate

(anionic polysaccharides) forms a network of heterogeneous gels depending on the alginate

concentration, Ca2+ concentration and its composition.56 The solution-state behaviors of EPSs differ

substantially in their 3D structures, degree of branching, conformational flexibility of backbone, charge

and their ability to interact with proteins.22-25 Besides these factors, conformational flexibilities of

monosaccharides and the nature of glycosidic bonds in their repeat units also greatly contribute to

solution-state behavior of EPSs.10, 57 Based on the Φ and Ψ angles analyses derived from best-fit

HePS-1 structure, we showed that HePS-1 has significant flexibility due to presence of Galf residues

and three (1→6)-linkages in repeating units, and in dilute solution at 1 mg/ml it behaves as a random

coil-like structure. The superimpositions showed that the five best-fit structures exhibited helix-like and

random coil-like structures (Figure 4B and D). In their natural environment, the structural flexibility

and the coil-like structures of HePSs may protect the bacteria against environmental stress. In protein-

free system (in vitro), we speculate that the coil-like structure may not contribute significantly to

viscosity because it may occupy less volume compared to its extended chain conformation. In protein-

containing system, the coil-like structure may stabilize the system by forming optimal surface contacts

with proteins, and thus it can be used for various applications.

21

Table 2. Summary of Φ and Ψ angles in HePS-1 structures

O5-C1-O1-C‛X C1-O1- C‛X-C‛X−1

Glycosidic Linkage Φ (°) Ψ (°)

α-D-Galp-(1→3)-β-L-Rhap 76 ± 29 94 ± 33

β-L-Rhap-(1→4)-β-D-Glcp

76

58 ± 40

93

111 ± 33

β-D-Glcp-(1→6)-β-D-Galp

66

−77 ± 33

110

−160 ± 40

β-D-Galp-(1→6)-α-D-Galp

−71

−86 ± 42

−160

−149 ± 40

β-D-Glcp-(1→6)-α-D-Galf

−75

−74 ± 43

−160

−174 ± 40

α-D-Galf -(1→6)-β-D-Glcp

−73

161 ± 43

175

−160

−159 ± 40

−160

The first Φ and Ψ values are from the best-fit HePS-1 models, and the second

values are from the starting HePS-1 model (Materials and Methods).

22

Figure 5. Plots of glycosidic dihedral angles for the HePS-1. Phi (Φ) and Psi (Ψ) angles, in all cases, are shown by gray circles. The dihedral angles of glycosidic linkages of disaccharides (1→3) are defined as Φ = O5‛-C1‛-O1-C3 and Ψ = C1‛-O1-C3-C2. For glycosidic linkages (1→4), the atoms defining Φ are O5‛-C1‛-O1-C4 and Ψ are C1‛-O1-C4-C3. For glycosidic linkages (1→6), the atoms defining Φ are O5‛-C1‛-O1-C6 and Ψ are C1‛-O1-C6-C5.

Ab initio scattering modeling of HePS-2 and HePS-3. Constrained scattering modeling

was not done for HePS-2 and HePS-3 because of their large molecular size which hampered

the generation of sufficient conformational models to be able to determine the best-fit

23

models. Instead ab initio models for HePS-2 and HePS-3 were generated from their X-ray scattering

curves, and in order to present overall shape envelopes for both of these HePSs in solution.

For HePS-2, ab initio scattering modeling analyses showed good agreements with SAXS data. The

20 HePS-2 models gave an average chi2 (χ2) of 0.46 ± 0.02 (where χ2 is the discrepancy between

theoretical and experimental curves) and the best-fit model gave χ2 of 0.43. The best-fit model gave RG

of 11.56 nm and L of 39.6 nm, in good agreement with the experimental GNOM RG of 11.49 ± 0.25

nm and L of 38.0 nm (Table 1). The P(r) curve for the best-fit model and their r values agree

with experimental data (Figure 6, inset). The visual agreement between the experimental

I(Q) curve and the modeled I(Q) curve was excellent (Figure 6).

Figure 6. The ab initio solution scattering curve fit for the best-fit HePS-2 model. The data on the left are the modeled scattering I(Q) curve for the best-fit ab initio HePS-2 model (blue line) fitted to the experimental scattering data (black circles). The best-fit (blue spheres) and the average of 20 models indicated by light gray surface are shown to the right. The maximum lengths L of best-fit ab initio model (blue) and experimental L are compared in the P(r) curve (inset).

For HePS-3, ab initio modeling analyses showed similar good agreements with the experimental

SAXS data. The 20 HePS-3 models generated from its X-ray scattering data gave an average χ2 of 0.37

± 0.01 and the best-fit model gave χ2 value of 0.34. The best-fit modeled curve I(Q) showed excellent

24

visual agreement with the experimental I(Q) (Figure 7). The best-fit model gave RG of 11.63 nm

and maximum length L of 39 nm, in good agreement with the experimental RG of 11.44 ± 0.1 nm

and L of 41.8 nm (Table 1). The P(r) curve for the best-fit model and their r values agree well

with those seen experimentally (Figure 7, inset). In conclusion, the best-fit ab initio

scattering models for both HePS-2 and HePS-3 also show compact structures.

Figure 7. The ab initio X-ray scattering curve fit for the best-fit HePS-3 model. The data on the left are the modeled scattering I(Q) curve for the best-fit ab initio HePS-3 model (blue line) fitted to the experimental scattering data (black circles). The best-fit (blue spheres) and the average of 20 models indicated by light gray surface are shown to the right. The maximum lengths L of best-fit ab initio model (blue) and experimental L are compared in the P(r) curve (inset).

Dynamic light scattering (DLS) of HePS-1, HePS-2 and HePS-3. DLS assesses the size of

macromolecules in solution by following their diffusion rates under Brownian motion. In DLS

experiments, sizes of the

25

three HePSs (0.51 mg/mL) in 10 mM sodium citrate buffer pH 4.0 were analyzed by using particle

sizing software. HePS-1 gave a single peak with average hydrodynamic diameter (dH) of 29.9 ± 0.2

nm (Figure 8A), which is smaller than the maximum length L determined by SAXS (Table 1). The dH

Figure 8. Dynamic light scattering data analyses of HePS-1. HePS-2 and HePS-3. Hydrodynamic diameter (dH) values of 29.7 nm for HePS-1 (A), 33.6 nm for HePS-2 (B) and 36.1 nm for HePS-3 (C). The DLS experiments were performed in 10 mM citrate buffer pH 4.0.

of HePS-2 and HePS-3 were determined to be 34.1 ± 0.3 and 35.6 ± 0.5 nm, respectively (Figure 8B

and C). In both Figure 8B and C a minor peak appeared at high dH values of 240260 nm which

26

could be attributable to dust particles (not shown). These dH values of HePS-2 and HePS-3 also

show slight deviation from their scattering lengths (Table 1), and reflect the difference in accuracy

of the two techniques. In DLS, dH is the diameter of the hard sphere that diffuses at the same rate as

the molecule being measured. Like X-ray data, these dH values are also much smaller than those

measured for the extended HePS-1, HePS-2 and HePS-3 models, thus supporting the X-ray scattering

data.

CONCLUSIONS

Knowledge on molecular structures of HePSs is essential to understand their function and the

relationship between their structure, and their physical and rheological properties. Despite information

on primary structures of oligosaccharide repeats as determined by NMR spectrometry, molecular

structures for full-length, intact HePSs have up to now not been reported. In the present study, we used

SAXS and DLS in combination with constrained scattering modeling to characterize molecular

structures of three LAB-HePSs in solution. SAXS data show that the experimental RG, RXS and

maximum lengths L values of HePS-1, HePS-2 and HePS-3 do not show linear relationship with the

corresponding modeled values calculated for their extended chain lengths, suggesting that these HePSs

exhibit compact structures of random coil-like conformation in solution. Here a constrained modeling

approach previously used to determine molecular structures for multi-domain proteins and more

recently for anionic oligosaccharides,3134 was extended to study full-length, large HePS. Constrained

scattering modeling determines a 3D medium-resolution structure and represents a family of best-fit

structures that best accounts for the experimental scattering data, suited also to derive structural

information on conformational flexibility of HePSs. Constrained scattering modeling of HePS-1

27

validated the experimental scattering data analysis. From analyses of Φ and Ψ angles of the best-fit

models, the findings suggest that consecutive (1→6)-linkages with Galf residues are more flexible and

produce progressive bents in the backbone tending to adopt non-uniform local helical and random coil-

like structures.

Ab initio scattering modeling analyses showed that HePS-2 and HePS-3 also adopt a coil-like

conformation in solution. Although these HePSs differ in their molecular sizes, their maximum length

values in solution were very close to each other, showing that the high molecular weight HePS has

more compact conformation than the smaller one, and this could be explained by high degree of

branching and flexibility of the backbone. The bends in the HePS-1 structures arise through kinks

produced by Galf in its chain. In the case of anionic polysaccharides interactions with proteins, it has

already been established that kink structure provided optimal electrostatic and hydrogen bond contacts

with proteins.58 In terms of interactions of HePSs with proteins, the bent structures may facilitate

optimal contact formation between HePSs and proteins, and may form stable complexes.

ASSOCIATED CONTENT

Supporting information

Size exclusion chromatography of HePS-1, HePS-2 and HePS-3

AUTHOR INFORMATION

Corresponding Author Department of Micro- and Nanotechnology, DTU, Ørsteds Plads, building 423, DK-2800 Kgs. Lyngby,

Denmark. phone: +4545258144. *Email: kral@nanotech.dtu.dk.

ACKNOWLEDGEMENTS

We thank Karina Jansen and Lotte Nielsen for their valuable technical assistance. This project was

funded by the Danish Research Council for Independent Research | Technical and Production Sciences

28

to the project “HEXPIN. Associative interactions between exopolysaccharides from lactic acid baceria

and milk proteins. Gaining insights deployable in design and optimised food texture”. We also thank

Prof. Luc de Vuyst, Vrije Universiteit of Brussel, Belgium, for providing the Streptococcus

thermophilus EU20 strain. The MAXIV synchrotron and the I911-SAXS beamline crew are

acknowledged for beam-time and support during data-collection. The research presented has received

funding from BioStruct-X and DANSCATT (the Danish agency and for Science, Technology and

Innovation).

REFERENCES

(1) Donot, F.; Fontana, A.; Baccou, J. C.; Schorr-Galindo, S. Carbohydr. Polym. 2012, 87, 951–962.

(2) Moscovici, M . Front. Microbiol. 2015, 6, 1-11.

(3) Dertli, E.; Colquhoun, I. J.; Gunning, A. P.; Bongaerts, R. J.; Le Gall, G.; Bonev, B. B.; Mayer,

M. J.; Narbad, A. J. Biol. Chem. 2013, 288, 31938-31951.

(4) Laws, A.; Gu, Y.; Marshall, V. Biotechnol. Adv. 2001, 19, 597–625.

(5) Torino, M. I.; Font de Valdez, G.; Mozzi, F. Front. Microbiol. 2015, 6, 834..

(6) Ganzle, M.; Michael, G.; Schwab, C. Exopolysaccharide production by intestinal lactobacilli. In:

Probiotics and prebiotics: Scientific aspects. Tannock, G. W (eds.). Norfolk, Caister Academic Press,

2005; pp 83-96.

(7) Schmid, J.; Sieber, V.; Rehm, B. Front. in Microbiol. 2015, 6, 1–24.

29

(8) Górska-Frączek, S.; Sandström, C.; Kenne, L.; Rybka, J.; Strus, M.; Heczko, P.; Gamian, A.

Carbohydr. Res. 2011, 346, 2926-2932.

(9) Ruas-Madiedo, P.; de los Reyes-Gavilán, C. G. J. Dairy Sci. 2005, 88, 843–856.

(10) Caggianiello, G.; Kleerebezem, M.; Spano, G. Appl. Microbiol. Biotechnol. 2016, 100, 3877-86.

(11) Badel, S.; Bernardi, T.; Michaud, P. Biotechnol. Adv. 2011, 29, 54–66.

(12) Ravishankar Rai V, Jamuna A. Bai (2014). Beneficial Microbes in Fermented and Functional

Foods. December 17, by CRC Press Reference - 600 Pages - 93 B

(13) Patel, A.; Prajapati, J. B. Adv. Dairy Res. 2013, 1, 1-7.

(14) Levrat-Verny, M. A.; Behr, S.; Mustad, V.; Rémésy, C.; Demigné, C. J. Nutr. 2000, 130, 243-

248.

(15) Maeda, H.; Zhu, X.; Suzuki, S.; Suzuki, K.; Kitamura, S. J. Agric. Food Chem. 2004, 52, 5533-

5538.

(16) Kim, Y.; Oh, S.; Kim, S. H. Biochem. Biophys. Res. Commun. 2009, 379, 324-329.

(17) Gentès, M-C.; St-Gelais, D.; Turgeon, S. L. Dairy Sci. Technol. 2013, 93, 255–271.

(18) Salazar, N.; Gueimonde, M.; de Los Reyes-Gavilán, C. G.; Ruas-Madiedo, P. Crit. Rev. Food

Sci. Nutr. 2016, 56,1440-1453.

(19) Doleyres, Y. L.; Schaub, L.; Lacroix, C. J. Dairy Sci. 2005, 88, 4146–4156.

(20) Ruas-Madiedo, P.; Tuinier, R.; Kanning, M.; Zoon, P. Int. Dairy J. 2002, 12, 689–695.

30

(21) Faber, E. J.; Zoon, P.; Kamerling, J. P.; Vliegenthart, J. F. G. Carbohydr. Res. 1998, 310, 269–

276.

(22) Tuinier, R.; Ten Grotenhuis, E.; Holt, C.; Timmins, P. A.; De Kruif, C. G. Phys. Rev. E 1999,

60, 848–856.

(23) Looijesteijn, P. J.; Schols, H. A.; Voragen, A. G. J.; Zoon, P. Biopolymers 2001, 59, 160–166.

(24) Girard, M.; Schaffer-Lequart, C. Int. Dairy J. 2007, 17, 666–673.

(25) Turgeon, S. L.; Plesca, V. Proceedings of the 5th International Symposium on Food Rheology

and Structure, Zurich, Switzerland, 2009; pp 416–419.

(26) Whistler, R. L., BeMiller, J. N. Carbohydrate chemistry for food scientists. Eagan Press, St

Paul, Minnesota, 1997; pp 117-151.

(27) Tuinier, R.; Zoon, P.; Cohen-Stuart, M. A.; Fleer, G. J.; de Kruif, C. G. Biopolymers 1999, 50,

641–646.

(28) Marshall, V. M.; Dunn, H.; Elvin, M.; McLay, N.; Gu, Y.; Laws, A. P. Carbohydr. Res. 2001b,

331, 413–422.

(29) Landersjo, C.; Yang, Z.; Huttunen, E.; Widmalm, G. Biomacromolecules 2002, 3, 880-884.

(30) Goh, K. K.; Haisman, D. R.; Singh, H. Appl. Microbiol. Biotechnol. 2005, 67, 202-208.

(31) Perkins, S. J.; Okemefuna, A. I.; Fernando, A. N.; Bonner, A.; Gilbert, H. E.; Furtado, P. B.

Methods Cell Biol. 2008, 84, 375–423.

31

(32) Perkins, S. J.; Okemefuna, A. I.; Nan, R.; Li, K.; Bonner, A. J. R. Soc., Interface 2009, 6, S679–

S696.

(33) Khan, S.; Fung, K. W.; Rodriguez, E.; Patel, R.; Gor, J.; Mulloy, B.; Perkins, S. J. J. Biol. Chem.

2013, 288, 27737-27751.

(34) Khan, S.; Gor, J.; Mulloy, B.; Perkins, S. J. J. Mol. Biol. 2010, 395, 504-521.

(35) Gruter, M.; Leeflang, B. R.; Kuiper, J.; Kamerling, J. P.; Vliegenthart, J. F. Carbohydr. Res.

1993, 239, 209-226.

(36) DuBois, M.; Gilles, K. A.; Hamilton, J. K.; Rebers, P. A.; Smith, F. Anal. Chem. 1956, 28, 350–

356.

(37) Labrador, A.; Cerenius, Y.; Svensson, C.; Theodor, K.; Plivelic, T. J. Phys.: Conf. Ser. 2013,

425, 072019.

(38) Petoukhov, M. V.; Franke, D.; Shkumatov, A. V.; Tria, G.; Kikhney, A. G.; Gajda, M.; Gorba,

C.; Mertens, H. D. T.; Konarev, P. V.; Svergun, D. I. J. Appl. Crystallogr. 2012, 45, 342-350.

(39) Glatter, O.; Kratky, O. (Editors): Small-Angle X-ray Scattering. Academic Press, New York, NY,

1982.

(40) Semenyuk, A. V.; Svergun, D. I. J. Appl. Crystallogr. 1991, 24, 537–540.

(41) Bohne,A.; Lang, E.; von der Lieth, C. W. J. Mol. Model. 1998, 4, 33-43.

(42) Bohne,A.; Lang, E.; von der Lieth, C.W. Bioinformatics 1999, 15, 767-768

(43) Mayo, S. L.; Olafson, B. D.; Goddard, W. A. J. Phys. Chem. 1990, 94, 8897–8909.

32

(44) Wright, D. W.; Perkins, S. J. J. Appl. Crystallogr. 2015, 48, 953-961

(45) Okemefuna, A. I.; Nan, R.; Gor, J.; Perkins, S. J. J. Mol. Biol. 2009, 391, 98–118.

(46) Ashton, A. W.; Boehm, M. K.; Gallimore, J. R.; Pepys, M. B.; Perkins, S. J. J. Mol. Biol. 1997,

272, 408–422

(47) Perkins, S. J.; Weiss, H. J. Mol. Biol. 1983, 168, 847–866.

(48) Franke, D.; Svergun, D. I. J. Appl. Crystallogr. 2009, 42, 342–346.

(49) Volkov, V. V.; Svergun, D. I. J. Appl. Crystallogr. 2003, 36, 860–864.

(50) Petoukhov, M. V.; Svergun, D. I. Biophys. J. 2005, 89, 1237–1250.

(51) Li, Y.; Huang, Q.; Shi, T. J. Phys. Chem. B 2006, 110, 23502-23506.

(52) Cheng, G.; Melnichenko, Y. B.; Wignall, G. D.; Hua, F.; Hong, K.; Mays, J. W.

Macromolecules 2008, 41, 9831-9836.

(53) Sánchez-Medina, I.; Frank, M.; von der Lieth, C. W.; Kamerling, J. P. Org. Biomol.

Chem. 2009, 7, 280-287.

(54) Faber, E. J.; van Kuik, J. A.; Kamerling, J. P. Vliegenthart, J. F. Biopolymers 2002, 63, 66-76.

(55) Dogsa, I.; Kriechbaum, M.; Stopar, D.; Laggner, P.Structure of bacterial extracellular polymeric

substances at different pH values as determined by SAXS. Biophys. J. 2005, 89, 2711– 2720.

33

(56) Stokke, B. T.; Draget, K. I.; Smidsrod, O.; Yuguchi, Y.; Urakawa, H.; Kajiwara,

K. Macromolecules 2000, 33, 1853−1863. (57) Hallemeersch, I., De Baets, S., Vandamme, E. J.

Biopolymers 2002, 5, 407-429.

(58) Raman, R., Venkataraman, G., Ernst, S., Sasisekharan, V.; Sasisekharan, R. Proc. Natl. Acad.

Sci. U.S.A. 2003, 100, 2357–2362.

34

Table of Contents graphic (TOC)